Circulating tumor cell enumeration, biomarker analyses, and kinetics in patients with colorectal cancer and other GI malignancies

“Liquid biopsies” are revolutionizing cancer care. The broad term often refers to identification of circulating tumor DNA (ctDNA) and circulating tumor cells (CTCs) in individuals with cancer. Most of the work pertaining to liquid biopsies has been about ctDNA. Data pertaining to work on CTCs as the other kind of liquid biopsies is scarce. Here we embarked on an effort to not only show feasibility of doing CTC enumeration and biomarker analyses, but also serial kinetics of CTCs in response to therapy in patients with colorectal cancer and other GI malignancies. This was a prospective study with collection and analyses of circulating tumor cells (CTCs) done in real-time in patients with colorectal cancer and other GI malignancies. We chose to identify CTCs using two different methodologies/platforms: RareCyte platform which relies on sepration based on the density of CTCs, and Parsortix platform which relies on the size/compressibility of CTCs. The study was reviewed by the institutional review board (IRB), and all patients (n = 59) signed a written informed consent before collecting blood samples at the University of Iowa Holden Comprehensive Cancer Center. Blood samples (4 tubes: 2 tubes for Parsortix system and 2 tubes for Rarecycte system) were collected from each patient at each time point. Each tube was subjected to CTC analysis and results averaged (Total = 150 draws; 281 test results) from all the patients. In addition to identification and enumeration of CTCs using standard definitions of nuclear/epithelial markers, additional biomarkers studied included PD-L1, HER2, EGFR, Ki-67, and where adequate confirmation with panel based single cell sequencing was also performed. Cell cultures were attempted for the parsortix collections. Patients with colorectal cancer (CRC) had the highest number of CTCs (Mean: 15.8; Median: 7.5). This was in sharp contrast to the number of CTCs in patients with pancreatic cancer (PC) (Mean: 4.2; Median: 3). At least a third of patients with CRC had at least 3 CTCs. Another 17.4% and 11.4% had at least 10, and at least 20 CTCs in CRC; while no patients with PC had more than 20 CTCs. In addition, we detected a significantly lower number of CTCs from patients receiving treatment versus patients not receiving treatment (Median 2.7 versus 0.7). Striking differences were also observed in CTCs for those with untreated/progressive disease versus those who had responding/stable disease (Median of 2.7 versus 0). Biomarker analyses on 4 markers HER2/PD-L1/Ki-67/EGFR were feasible when CTCs were detected. There was overall agreement on CTCs detected on either platform. Not suprisingly, the agreement was higher when more CTCs were detectable. Where available, single cell sequencing data was concordant with clonal aberrations on ctDNA/tissue NGS sequencing. This is one of the largest prospective endeavors in GI cancers showing feasiblity and value of capturing and analyzing CTCs in a prospective fashion using two very different platforms. Timing is key to CTC based protocols and analyses. Varied downstream capabilities of different platforms and the fact that they are capturing an overlapping but different bucket of CTCs lends pros and cons to each approach.